Distributed Kalman filtering for sensor network with balanced topology

Abstract In this paper, we study the distributed Kalman filtering for sensor network with mild assumption on communication topology and local observability. To this end, a new peer-to-peer distributed Kalman filtering is proposed, where each sensor communicates with its connected neighbors to achieve average consensus on weighted measurements and inverse-covariance matrices. Then, a consensus strategy is introduced to reduce the error produced by embedded dynamic consensus method. In addition, the convergence and steady-state performance of the proposed algorithm are also investigated, and we prove rigorously that the biased estimates are bounded and controllable. Numerical simulations validate the theoretical contributions of this paper.